"THE DEVELOPMENT OF SCIENTIFICNETWORKS AND MULTINATIONAL

RESEARCH PROJECTS: WHOPLAYS THE KEY ROLE?"

by

A. DE MEYER*W. SAAB**

andJ. TABET***

94/12/TM

* Professor of Technology Management at INSEAD, Boulevard de Constance, 77305Fontainebleau Cedex, France.

** Research Associate, at INSEAD, Boulevard de Constance, 77305 Fontainebleau Cedex,France.

*** Research Associate, at INSEAD, Boulevard de Constance, 77305 Fontainebleau Cede;France.

A working paper in the INSEAD Working Paper Series is intended as a means whereby afaculty researcher's thoughts and findings may be communicated to interested readers. Thepaper should be considered preliminary in nature and may require revision.

The Development of Scientific Networksand Multinational Research Projects:

Who Plays the Key Role ?

A. De Meyer(*)W. Saab and J. Tabet 0')

I - Introduction

Science is a matter of competitive co-operation (Hull [4]). Co operation and collaborationamong scientific researchers have always occurred, either initiated spontaneously by someresearchers, or encouraged by research organisations who believe that collaborative work ismore productive than that of an individual work. This organised collaboration is particularlyencouraged by public research authorities in their sponsoring of multinational researchprogrammes, as a response to society's major problems. In such a situation, rapid mobilisationof scientific resources is needed, and the existence of scientific networks is the cornerstone ofefficiency.

The aim of the research presented here is to understand more clearly the factors that encourageinternational projects to take place, and influence the creation or extension of Europeanscientific networks in particular.

The study focused on two European scientific communities. A questionnaire was designed,based on a literature review and on preliminary interviews with scientists. The analysis of theresults provided us with an understanding concerning the process of information search byresearchers, and some other behavioural aspects of researchers.

We have also analysed the communication process among scientists, which reveals theexistence of scientific networks. It appears that some researchers play a key role either inexchanging information inside and outside their research organisation, or in communicatingwith other researchers. We have attempted to identify these people, to determine theirdifferentiating characteristics, and to confirm their role in the constitution of internationalprojects. Our contribution has been to analyse the role of these communicators in an importantand specific environment: The European multinational communities of scientists.

Finally, some managerial implications concerning the process of information search and thebuilding of European scientific networks were drawn from these results.

() Professor of Technology Management, INSEAD, Boulevard de Constance 77305 Fontainebleau CedexFrance.

(**) Research Associates at the European Institute of Business Administration (INSEAD). TechnologyManagement Department. Boulevard de Constance 77305 Fontainebleau France.

II - Literature review

The importance of communication for researchers' performance has been a recurrent theme inthe literature.

According to Allen [1,6], the performance of an R&D project group focusing on technologicaldevelopments is not necessarily correlated with communication inside the group. It is rathercorrelated with the density and diversity of communication with researchers from other groupsinside or outside their laboratory. Allen has also established that the communication outside theorganisation proceeds through a few performing communicators called TechnologicalGatekeepers. He has stressed the role of these technological gatekeepers in the disseminationof external scientific information inside the organisation.

2

Katz [5,6,7] distinguishes three kinds of R&D projects :Research oriented projects: They are universally defined because fundamental scientists, asopposed to technologists, speak a universal language that is understood by all scientists.Therefore the performance of these projects is better when all the participants in a researchproject communicate with external sources of information.Product and process development projects: They are strongly local because they are based ontechnologies that are specific to a company or an industry. Therefore the best sources ofinformation for these projects are inside the organisation. Nevertheless, external informationand new technological advances are exceedingly important for development projects.Technological gatekeepers are needed to bridge the gap between local development work andthe larger technological world outside the organisation. However, direct individual externalcontacts are not correlated with higher performance. Efficient external contacts are those thatproceed through the technological gatekeepers.Product and process modification or adaptation: They are even more local and deal with moremature technologies. The gatekeepers' role is not as critical as in development work.

The importance of communication between scientists and the role of the gatekeepers have beenemphasised or criticised by several other authors (Epton [2], Keller and Holland [8], Persson[9], Tushman [11]).

It is important to note two important differences between the environments described in thoseprevious works and ours:

1. Allen's work was focused on the technology field as opposed the scientific one. This iswhere it differs significantly from this research, which was focused on science andfundamental research.

2. The scientific communities studied here are multinational, since they are composed ofresearchers from different European countries.

The present study focuses on communication over large geographical distances and betweenresearchers belonging to different organisations.According to Tomlin [10], researchers will communicate over long distances only if theybelieve that they have strong reasons to do so. Inter-organisational links, as well as superior-subordinate links, can bridge considerable physical distances. This is important in the case of aEuropean scientific community where researchers are located in different countries. Howells[3] addressed the advantages and disadvantages of concentrated versus dispersed patterns ofR&D. According to him, the location context of R&D remains a neglected but increasinglysignificant issue, with the growing internationalisation of R&D.

Vinck et al. [12] have analysed how new scientific communities are created in response topublic policy initiatives. They argue that these communities are formed through formal orinformal networks, according to the type of policy initiative. They describe these networks asflexible co-operation networks. They have described three mobilisation modes in the EuropeanPublic Health and Research Programmes. In type 1, an initiator comes with his already definedteam, and submits a proposal to the EC, which invites the other members that he has chosen toa preliminary meeting. Then, the presumed initiator assumes the leadership of the project. Intypes 2 and 3, mobilisation passes through the call-for-tender process, and the team leader who

3

submits the best proposal is in charge of seeing through the research project. In the third type ,the leader comes already with a complete unit, while in the second type, only the core of theteam is formed, and the complementary skills (or units) are formed subsequently with thesupport of the EC formal network. In the first two mobilisation modes, networks are originallyformed around a small number of teams (3-6), while in the third mode, networks are almostcompletely operational.

All the contributions mentioned here above addressed the following questions

1. How does communication proceeds in European scientific communities?2. What is the role played by the "high communicators" in these communities?3. How does effective collaboration proceed over long distances and different

countries?4. What are the consequences on the international research projects?

However these questions had never been addressed in such environments as EuropeanScientific Communities. This is what we did specifically in this study.

111 - Methodology

The project followed four steps:- Selection of two scientific communities on which the study was to be focused, and

interviews with researchers from each selected community;- Development of a questionnaire aimed at answering the questions raised by the literature

and interviews;- Distribution of the questionnaire to the communities' researchers;- Analysis of the results leading to reports.

4

1. Selection of the communities

For a good understanding of the role of communication and collaboration among scientists, itwas decided to study two scientific communities that have significantly differentcharacteristics. This was important for two reasons:

1. To avoid biases due to the idiosyncratic characteristics of one community.

2. To allow us to draw general lessons applicable to other communities from the analysisof differences and commonalities between two communities.

We finally chose the two following groups:

The AIDS community: Researchers working directly on or contributing to thedevelopment of a vaccine against AIDS;The Photo voltaic community: Researchers working on the conversion of the solarenergy into source of power.

Research into Aids is intense because of its crucial importance to society. It receivesconsiderable interest and support from national authorities, and is covered intensively by themedia. The potential prestige for any scientist who might make a discovery is very high. Thepotential commercial profits are considerable too. On the other hand, research on the photovoltaic energy has for a long time been driven by long-term goals, with no immediate industrialapplications or short-term commercial profits. It is driven by a vision: "A clean renewableenergy for the world". Although R&D on the photo voltaic energy started in the earlyseventies, it has received much less support and fewer subsidies than research into AIDS.

From a scientific point of view, research on AIDS relies essentially on bio-medical scienceswhich can be described as empirical, while photo voltaic research relies on physics that can bedescribed as theoretical. The specificities of each community along with the differencesbetween them, have justified their being chosen for this study. In this paper, we will focus moreon the commonalities than on the differences.

2. Interviews with scientists

Approximately ten interviews were conducted with senior scientists and laboratory managers ineach scientific community. This exploratory study helped us to design the questionnaire. A"pilot questionnaire" was first designed and shown to 34 researchers from both communities.The questionnaire was then modified according to their feed-back and a final draft wasrewritten in a way that minimised confusion and misunderstandings.

The interviews also allowed us to select, from all the issues raised in the literature, those whichwere relevant for the selected communities. As a matter of fact, the literature that inspired uswas based mainly on research in the United States. We wanted to check that the conceptsdescribed in this literature were also applicable in a European, multinational environment. Thiswas achieved through the interviews.

5

3. The questionnaire

a. ContentThe questionnaire was intended to provide data and information about European scientists'behaviour, in order to facilitate a better understanding of the creation and development ofEuropean R&D networks. It was focused on a detailed analysis of the means and frequency ofcommunication between scientists.

The questionnaire addressed several other issues such as the means of carrying outbibliographic search, collaboration with other research teams, perception of the role of the EC,and asked a series of questions aimed at exploring the cultural differences between researchersfrom different countries. These issues however are not discussed in the present paper.

b. Scientific Networks as Communication NetworksA network is a structure that links researchers. Its functioning is reflected by thecommunication between these researchers. We have analysed this network by means ofanalysing the communication in greater depth. In particular, we have analysed the means, thetype and the frequency of the communication between researchers.

In order to gauge this communication activity, we adopted a dual approach, based on acombination of questions about frequency and means of communication.

First, researchers were asked to rate their frequency of communication on a seven point scale,by answering the following three questions':

How often do you communicate with other researchers:1. inside your organisation?2. outside your organisation (but in the same country)?3. in another country?

These questions also allowed us to distinguish between researchers who communicated onlywith fellows within their organisation from those who communicated frequently outside theirorganisation.

In order to differentiate formal communication (scheduled) from informal communication (notscheduled), researchers were also asked to define their means of communication and theirrelative frequency of use. Informal meetings and the telephone were associated with informalcommunication. Formal meetings and correspondence were associated with formalcommunication.

c. Tyne of auestionsThree types of questions were asked: One hundred and six "I to 7" scale questions, Six "Yes -No" questions and five Multiple choice questions.

1 Each question was to be rated on a 1-7 scale.

6

Most of the information yielded by the questionnaire stems for a self-rating by the researchersof their behaviour (on a 1-7 scale). This method may seem subjective. The experience showsthat the absolute values obtained by self-rating can be biased. However the results of self-ratingreveal the correct trends within a group. An individual can find it difficult to give an absoluteevaluation of his own behaviour but he generally knows how he compares with other persons.Since comparisons between individuals (or groups) and correlations between differentbehavioural aspects of individuals are what was being investigated in this study, self-rating wasa suitable method for it. Moreover, self-rating allowed us to obtain the information in a muchmore concise way than direct questions would have done. This allowed us to keep thequestionnaire to a reasonable length. Indeed, a compromise had to be stuck between the levelof detail and the total length of the questionnaire.

4. Administering the questionnaire

For the photo voltaic community, the database of potential respondents consisted of the list ofall the participants at the eleventh Photo Voltaic Solar Energy Conference held in 1992 inLisbon (Portugal). This conference is the most important one held in Europe in this particularfield. It is reasonable to say that the participants at this conference represent quite accuratelythe European photo voltaic community. Approximately 900 researchers were selected in thisway.

Concerning the AIDS community, the mailing list was built differently. Three complementarysources helped to constitute a list of researchers working specifically on the development ofthe vaccine (World Health Organisation, "Agence Nationale de la Recherche contre le SIDA"and "Institut Pasteur" in France). Personal contacts in other European research instituteshelped in adding to these lists researchers working specifically on AIDS. All the lists wereaggregated to produce a final mailing list of 1800 researchers. This mailing list includesvirtually all members of the French AIDS community, and a representative sample of the otherEuropean AIDS communities.

The questionnaire was therefore sent by mail to 2700 researchers from the two communities.

5. Comments

A few comments must be made in order to explain potential biases or problems encounteredwith our methodology.

The evaluation of the individual performance of researchers, as well as the performanceof a research unit, are important parameters needed for this study. However, askingdirect questions about researchers' performance would give them the impression of beingaudited, or judged according to their output. For these reasons, questions aboutperformance were limited to the number of papers authored (or co-authored), and thenumber of projects carried out at the same time. An estimate of the productivity isreflected in the ratio of the number of papers published, divided by the author's years ofexperience (Papers/Year).

7

The evaluation of researchers by the quantity of papers they have published, has beencriticised by the research organisations themselves. Obviously, it has led researchers topublish several papers with nearly the same content in different journals. Therefore, someresearch organisations (e.g. "Centre National de la Recherche Scientifique" in France)also evaluate their researchers according to the quality of their papers. However, theevaluation of the quality of a paper is somewhat subjective. At the same time, manymanagers in research organisations point out that the evaluation of researchers by thequantity of papers is well adapted in most cases. They argue that it is an objectivemeasurement; and also that the researchers' productivity in terms of quantity of papers iscertainly the expression of a contribution to the life of their scientific community thatdeserves a recognition. Without entering the debate, we have chosen this evaluation ofthe productivity of the researchers because it was the best objective productivityappraisal, that could be monitored through a questionnaire.

Apart from the classical problems with questionnaires, there is the problem of addressingan international population with a questionnaire in English. We have chosen not totranslate the questionnaire for two reasons:1. We assume that researchers understand English, which has become the

international scientific language.2. Translations in several European languages would have introduced biases into the

questionnaire.

There are inherent biases to this type of study. The questionnaires have been sentrandomly to a great number of researchers. One can reasonably argue that the proportionwho answered the questionnaire (14% of the whole targeted population) has differentcharacteristics different from the average researcher. It is very likely, for instance, thattheir knowledge of English is better than average, since the questionnaire was written inEnglish. While it was impossible to avoid these biases with our research design, theymust be kept in mind. Interpretation of the results and their general application to thewhole scientific community must be done carefully.

Some of the results presented were obtained either by comparing the average characteristics oftwo groups, or by testing linear correlations between some characteristics and a variable. Forall correlations or comparisons, a 95 % level of significance was taken in account.

IV - Results

1. Description of the samples

A total of 373 answers were received, and were fairly representative of the two communities inEurope (return rate=14%). 136 came from the Photo voltaic community. 237 came from theAIDS community. Characteristics of the samples were as follows:

1. Nationality: (Table I) 28 different nationalities were represented in the sample.

2. Institution: (Figure 1) Researchers from the AIDS community came from 4 types ofinstitution: Universities (29%), Research centres (52%), Industry (5.5%), Hospitals (35%).

8

Researchers from the photo voltaic community came from Universities (32%), Researchcentres (45%) and Industry (23%).

3. Age and experience: (Figure 2) The average age of the researchers was 43 and 40 yearsrespectively for the AIDS and Photo voltaic communities. The average experience was 18and 15 years respectively.

4. Gender: In each of the two research communities, women represented almost 10% of thesample, which is proportional to their representation in these scientific communities.

5.Hierarchical level: (Figure 4) Researchers were divided into five hierarchical groups:- Level 1: Researcher who is still completing his doctoral thesis (8%).- Level 2: Junior research scientist with no subordinates (14%).- Level 3: Senior research scientist with at least 1 subordinate (23%).- Level 4: Head of group / division / unit (20%).- Level S,• Head of department or professor (35%).

6. International experience and linguistic abilities: Two new defined parameters have beenused in this study: the International Exposure and the Linguistic Index.

International exposure = number of years spent abroad * number of countries visited.Linguistic index = sum of levels of fluency for each language spoken

Most of the researchers had worked in at least one foreign country and had mastered Englishwell. The average researchers in both samples had 4 years of international exposure in 1 or 2countries. They spoke on average 3 languages: their mother tongue, English (level-6 on a 1 to7 scale, when it was not the mother tongue), and a third language. One must keep in mind thatthere is a bias in these results, since they are certainly not really representative of the wholecommunities. Indeed, as mentioned above, we suspect that the researchers who answered thequestionnaire have a better command of English than average and that they have had a higherinternational exposure.

2. Importance of Researchers' Exchange Programmes

As mentioned by Tomlin [12], scientists will communicate over long distances only if theybelieve that they have strong reasons to do so. But inter-organisational links, as well assuperior-subordinate links, can bridge considerable physical distances. For this reason, it isimportant to move scientists between laboratories on an exchange basis, either for short orlong periods. This is believed to play a highly effective role in the collaboration process.

The researchers' answers to the questionnaire support these statements :

81% of the researchers who answered the questionnaire have been sent to other organisationsduring their career. Almost all researchers (99%) are willing to participate in such exchangesprovided that the objective of the visit is relevant to their own research (a programme shouldbe agreed in advance), and their travel and general expenses are covered. For personal reasons,many researchers favour visits of less than one month.

9

According to the researchers, these exchanges contribute to the creation or development ofscientific networks through the building of personal relationships, the identification ofcomplementary skills, the exchange of valuable information, technology transfers and the jointwriting of papers.

The benefits of such exchange programmes are shown in Figure 4. Further comments will bemade on the results shown in Figure 4, in paragraph V.

3. Means and frequency of communication

a. MeansThe means used by the researchers to communicate inside and outside their organisation aredetailed in Table 2. We have distinguished communication outside the organisation withresearchers involved in a common project from communication with researchers not involvedin any common project.

"Informal meetings" is the most frequently used means of communication by the researchersinside their organisation. It is followed by "phone" calls, and then by "formal meetings".This is in line with observations made by Allen [1], showing the high frequency of informalcommunication between researchers.

In order to communicate with researchers located outside their organisation, researchers woulduse mainly "phone" calls, then in decreasing order of frequency: "formal meetings","informal meetings" and "correspondence". The most frequent use of phone calls isjustified by the distance separating researchers from different organisations. The use of formalmeetings is made necessary in order to coordinate efforts of researchers from differentorganisations, and to maintain a formal character to the collaboration among -differentorganisations.

It is interesting to note that some researchers communicate frequently with researchers fromother organisations not involved in any of their research projects. This point will be discussedin paragraph 4.

b. FrequencyTo obtain the total frequency of communication, we simply summed up the frequencies ofcommunication inside and outside the organisation, and abroad as mentioned in paragraph II-4.The result is represented in figure 5.

4. The Communication Champions

It has already been mentioned by Allen[1], that "technological gatekeepers" communicate moreoften, especially outside their organisation, than other researchers. However, as we pointed outbefore, our environment differs from the one described by Allen with respect to two factors.Our communities are multinational and they are concerned with science rather than technology.

Therefore, it was important to isolate researchers who communicated more often than othersboth inside and outside their organisation. We have distinguished two groups: one normallydistributed around a mean of 13, the other starting at 18, exponentially increasing to reach itsmaximum at 21. This second group includes people who communicate significantly more than

10

average with people inside their organisation, outside their organisation and in other countries,simultaneously. We have defined these as Communication Champions.

Is it realistic to say that researchers who communicate very often both inside and outside theirorganisation, make a special contribution to their scientific community?To answer this question, it seemed necessary to identify other characteristics that they shared,in order to have a better idea of who they were, and how they behaved. The analysis of otherinformation available from the answers to the questionnaire helped to highlight someparticularities, that are not exclusive to the Communication Champions, but are significantlycorrelated with them. These characteristics are described in Table 3. This analysis wasachieved either by testing linear correlations between different characteristics and thecommunication champions taken as a dummy variable called "Champ", or by comparingaverage characteristics of champions with those of non-champions.

a. BackgroundCommunication Champions have experience outside their native country. Their internationalexposure, on average, is significantly higher than that of other researchers. Perhaps as aconsequence, they are more fluent in foreign languages, with an above average linguisticindex as well. Their English language skills appear to be high, maybe as a result of highexposure to scientific literature, which is by and large written in English.

b. BehaviourCommunication Champions show some behavioural characteristics highly correlated with theway they communicate. Personal contacts outside their organisation play an important role intheir work. Personal networks are, for them, a major source of scientific information, afterbibliographic search. Communication Champions hold more informal meetings than average.They also hold formal meetings very often, but not significantly more than average.Communication Champions seem often to communicate with researchers not involved intheir projects. This is again a distinctive feature, since they communicate with otherresearchers without it being necessary for the job they are assigned to. It is the result of apersonal choice or initiative, and not a consequence of their participation in internationalprojects. In terms of behaviour, it is important to mention that Communication Champions aretime sensitive. They believe to a greater extent than other researchers, that "time is money",meaning that they have no time to waste. Time sensitivity was surprisingly not affected bycultural aspects, as one would have expected. Similarly, no difference was noticed between thetwo scientific communities. Finally, one of the aspects that caught our attention was the wayCommunication Champions rate other researchers involved in their projects. They rate highlytheir willingness to collaborate. This is not surprising if we assume that they have goodrelationship with all their colleagues. It also reflects their own willingness to collaborate withothers, which is also probably above average.

c. Hierarchical level and DerformanceA significant positive correlation was reported between the hierarchical level and the dummyvariable "Champ". Although position and age are positively correlated, age is less significantlycorrelated than the position itself, meaning that the development of these networks is not theresult of a natural cumulation of personal contacts with time. Since the CommunicationChampions each play an important role in their institution, it remains difficult to determinewhether their high hierarchical level results from a recognition of their competence by a fairsystem, or whether the fact of being of a high status makes it easier to establish permanent

11

contacts with the outside world. Champions' individual performance (in terms of productivityin scientific papers) is also higher than average. One should be careful in interpreting thisresult, since a positive correlation between position and performance has already beenobserved.

d. Role and contribution to the research communityThere is a significantly positive correlation between the participation in international projectsand the "Champ" dummy variable. Although the champions' role in designing such projects cannot be precisely determined, they certainly contribute to organising multinational teams asrequested by EC-sponsored or other European research projects. This is facilitated by theirhigh hierarchical level, which gives them a relative autonomy, which in turn can be expected toshorten the decision making process. Without any doubt, Communication Champions are theinterface between their organisations and other research institutions. They are at the centre of ascientific network, forging links with the outside, and also involving their colleagues. Theirconsiderable experience abroad, together with their personal relationships, contribute to theexchange of more valuable information, and the identification of complementary skills. Finally,international projects appeared to be positively correlated with individual productivity. Thiscan be understood in two different ways:

International projects accept only researchers who have a distinctive performance(But we can question the fairness of the selection process for internationalprojects).Researchers' performance is improved by their participation in internationalprojects, either at the individual level, or at the community level.

V - Discussion

The role of the Communication Champions in the development of international projects issummarised in figure 6.

Before the materialisation of scientific networks by an international project or internationalcollaboration, some latent networks of scientists exist.

The personal relationships formed between researchers are at the source of these latentnetworks. As has been shown by the answers to the questionnaire, the researchers' exchangesencourage the formation of personal relationships between researchers from differentorganisations or countries (Figure 4-a). To a lesser extent, they also reduce cultural gaps(Figure 4-b). They also allow researchers to identify complementary skills, which is very usefulfor further collaboration (Figure 4-c).These researchers' exchange may result from spontaneous initiatives between laboratories, orfrom institutional programs, e.g.the EC Human Capital and Mobility program.

Not all the personal relationships developed during researchers' stays in other laboratories willlead to fruitful collaboration in the future. We would argue however, that CommunicationChampions, more than other researchers, maintain the relationships formed in the past, becausethey are conscious of their potential usefulness. Indeed we have seen that CommunicationChampions communicate frequently with researchers who are not involved in their projects(Table 3). This is not the case for other researchers. By doing so, the Communication

12

Champions are at the hubs of latent scientific networks. They activate them when there is apossibility of an international project (see Vinck et al.[12]).

Once an international project has been officially set-up, the latent networks and collaborationsare formalised. Scheduled meetings are held at regular intervals in order to monitor theevolution of the research program. However, as has been mentioned by the researchersthemselves, the exchange of valuable information and the technology transfers (see table 4-eand 4-fi occur more during the informal contacts between researchers. These informal contactsoccur between the formal meetings in two ways:

Researchers exchange valuable information on the occasion of mutually agreedworking visits. They are likely to share valuable information more easily if apersonal relationship has previously been built between them. That is one reasonwhy the visits that allow researchers to build personal relationships should beencouraged;Researchers can communicate at a distance by telephone or fax.

It must be mentioned that, once a collaboration between laboratories has been formalised, mostof the researchers are prepared to visit other laboratories only if the visit is directly relevant totheir own research. They want the visit to be short, concrete and focused on the subject ofinterest to them. This is a different purpose than just establishing personal relationships.

The building of personal relationships among researchers should therefore be encouraged on along-term basis. It need not necessarily take place in the framework of formal collaborationsbetween laboratories. Many other possibilities exist such as graduate training periods, doctoraland post-doctoral stays.

VI - Conclusion

The global performance of a community is gauged by its scientific and technological progresswith regard to the problems it has to solve. It is not unrealistic to believe that the globalperformance, expected of an international project, exceeds the sum of the individualperformances of the participants. This is the rationale behind the idea of actively promotingmultinational projects. Our work gives a better understanding of the functioning ofinternational projects and some guidelines to promote them.

It has been shown that Communication Champions in scientific communities play a similar roleto technological gatekeepers in technology development. They communicate frequently withresearchers from other organisations or countries, even with researchers who are not involvedin their projects. They are therefore at the hubs of latent scientific networks. They activatethese latent networks when there is a call for proposals for international research projects.

These Communication Champions have certain characteristics. Many of these characteristicsare those of the technological gatekeepers: they communicate more than average (ofteninformally); they are at high hierarchical levels; they are effective. Other characteristics seemspecific to the European scientific environment: they have had a strong international exposure,and they have strong language skills (especially in English).

13

We think that the Communication Champions can be recognised by research managers thanksto the characteristics described above. This information can help them in setting up contactswith other laboratories, based on personal relationships. The building of personal relationshipsamong researchers from different organisations is important for two reasons:1. -These relationships are at the origin of latent scientific networks that can be formalised to

take on international projects;2. -They promote the exchange of valuable information between researchers who collaborate

on a given project.

Research managers should therefore encourage their researchers, on a long term basis, tospend periods in other laboratories in order to build personal relationships between them andother researchers.

However, when researchers' visits to other laboratories are made in the framework of aformalised project, they should focus on a relevant subject. The mission should be clearlydefined in terms of tasks and time table. Short stays seem to be more appropriate in such cases.

14

Table 1Distribution of Researchers by Nationali

France 146 Greece 1United Kingdom 57 USA 11Germany 48 Canada 3Belgium 10 Mica 3Spain 14 Latin America 2Portugal 2 Japan 1Italy 13 Middle East 4Netherlands 15 Asia 3Denmark 3 Russia 2Sweden 1 Eastern Europe 12Switzerland 12 Austria 2Norway 1 Photo voltaic 136

AIDS 237Total 373

Finland 3Eire 4

15

Table 2Means of communication used by the researchers

Means ofcommunication

Frequency ofuse inside theorganisation(1 to 7 scale)

Frequency of useoutside the

organisation (1 to 7scale), with researchersinvolved in a common

research project

Frequency of useoutside the organisation

(1 to 7 scale), withresearchers not

involved in a commonresearch project

Informal meetings 5.21 4.38 3.25Formal meetings 3.97 4.47 3.37

Phone 4.55 5.64 3.71E-Mail 1.81 2.04 1.61

Correspondence 2.73 4.20 3.06

16

Fiiure 1 Distribution of Researchers by Institution

17

Fiffure 2Distribution of Researchers by Age and Experience

18

Figure 3Distribution of researchers by hierarchical level

19

6 71 2 3 4 514. 45111,7 ...lot

761 2 3 4 5

1.240110,7=414t

70%

60%

50%

10%

30%

20%

10%

0%

Figure 4Benefits of Researchers Exchange Programmes

a- Building personal relationship b- Identifying complementaryskills

c- Reducing cultural gapbetween organisations

d- Exchanging morevaluable information

e- Implementing technology transfer f- Writing papers together

1.41014, 7-.Y Imame, 7-a lot

20

Table 3Characteristics of the Communication Champions

Correlations and Comparisons

As indicated in section VI, we tried to characterise the champions of communication by testing linear correlations betweendifferent parameters and "communication champions" taken as a dummy variable. (The variable called "champ" takes thevalue 1 if the individual is a communication champion and the value 0 otherwise).The following table shows the parameters of the most significant correlations obtained:

Table (3-a)

Personal Communication withInternational English contacts Informal researchers not involved in

exposure outside meetings their projectsCommunication F = 8.2 F = 8.2 F = 37.2 F = 14.9 F = 22.8

Champions p < .0045 p <.0043 p < .0001 p <.000I _p <.0001

Other characteristics were derived from comparisons of the average answers of champions and non-champions to thequestions indicated below (rated on a 1 to 7 scale). The average answers of the two groups to each question are given. Allof them are significantly different at a 95% level.

Table 3-b

Questions Champions Others

1. a- Regarding the researchers from other organisations participating in your

research project, could you rate their willingness to collaborate? (1 .4ow, 7=high)

5.97 5.59

1. b- Their actual participation? (1=low, ?Adel) 5.36 4.99

2. a- To what extent are you satisfied with the scientific communication you have

with other researchers within your department, in tent's of frequency?

5.50 4.48

(Inot satisfied, 7-satisfied)

6- In terms of quality? (1 –not satisfied, 7=satisfied) 5.59 4.95

3. Do you give priority to facts and figures or to interpretation and theory ? 3.20 3.63

(1.,Facts, 7=Theory)

4. Regarding scientific reputation, to what extent is it important to write single

author articles vs. co-author articles ? (1=Not important, 7=Very important)

2.56 2.95

5.To what extent does the following assumption apply to research environments:4.43 4.00

"Time is money" ? (1– Not true, 7...Very true)

21

Figure 5

3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Total frequency of communication

22

INTERMIONALEXPOSURE

PERSONALCONTACTS

OUTSIDE THEORGANISATION

COMMUNICATIONWITH PEOPLE NOTINVOLVED IN THESAME PROJECTS

INFORMALCOMMUNICATION

INTERNATIONALCOMMUNICATION

LINSUISTIOINDEX COMMUNICATION

CHAMPIONSINDIVIDUAL

PERFORMANCE

-POSITION

SCIENTIFICNETWORKS

IINTERNATIONAL'

PROJECTS

I

23

References

[1] T.J. Allen, Managing the flow of technology, MI.T. Press, Cambridge, MA (1977).

[2] S. R Epton, Ten years of R&D management-Some major themes: the role of communication in R&D,R&D Management 11, 4, 165-170 (1981).

[3] J. Howells, The location and organisation of research and development: new horizons, Research Policy 19,133-146 (1990).

[4]D. L. Hull, Science as process: an evolutionary account of the social and conceptual development of science,University of Chicago Press, Chicago (1988).

[5]R. Katz, The effects of group longevity on project communication and performance,Administrative Science Quaterly 27, 81-104 (1982)

[6]R. Katz and T.J. Allen, Investigating the Not Invented Here (NIH) syndrome, R&D Management 12, 1, 7-19 (1982).

[7]R. Katz and M. Tushman, Communication patterns, project performance and task characteristics: Anempirical evaluation and integration in an R&D setting, Organisational Behaviour and HumanPerformance 23, 139-162 (1979).

[8]R. T. Keller and W. E. Holland, Communicators and innovators in research and development organisations,Academy of Management Journal 26, 4, 742-749 (1983).

[9]0. Persson, Notes and comments Critical comments on the gatekeeper concept in science and technology,R&D Management 11, 1, 37-40 (1981).

[10]B. Tomlin, Inter-location technical communication in a geographically dispersed research organisation,R&D Management 11, 1, 19-23 (1981).

[11]M. L. Tushman, Managing communication networks in R&D laboratories, Sloan Management ReviewWinter 1979, 37-50 (1979).

[12]D. Vinck, B. Kahane, P. Laredo & J. B. Meyer, A network approach to studying research programs:mobilizing and co-ordinating public responses to HIV/AIDS, Technology Analysis & Strategic Management5, 1, 39-54 (1993).

24